Refining the Shortest Paths (RSP)

Yuri Niella & Hugo Flávio

2020-03-03

1. Preparing the data

1.1. Preliminary analysis using actel

Analysing acoustic telemetry data requires initial filtering to exclude misleading data (e.g. false detections, detections prior to release). To overcome this issue and ensure reliable results, the RSP toolkit operates in close relationship with the actel R package, which filters and invalidates flawed detections. Before getting started with RSP, you will have to download actel and filter your acoustic data. Please click here for more information and to download actel.

To start using RSP, you can run the simplest actel analysis with the function explore. You can find more about how organize your data and run this preliminary analysis in actel’s manual pages (run ?actel after loading the library).

It is important that you save the output of the actel function, so you can later on use it to calculate the RSP. e.g.:

library(actel)
filtered_data <- explore(tz = "Europe/Copenhagen")

1.2. Creating and exporting a raster file from the study area in QGIS

After filtering your acoustic data, you will need a raster file from your study area defining the water and land limits. This file will be used for estimating the shortest paths in water between consecutive acoustic detections, using a least-cost analysis of constrained random walks. The values of the raster cells must comprise zeros (water) or ones (land). Depending on the size of your study area, a resolution of 0.0001 (latitude ° x longitude °) is suggested for more accurate estimations (especially for sites with very narrow channels). Please see the following steps for generating and exporting a raster file from you study site in the QGIS software:

  1. Load a good-resolution shapefile of your study region into QGIS:

  1. Zoom into your study site and create an overlapping polygon that encompasses the entire area,

  1. Clip your study site from the shapefile using “Vector > Geoprocessing Tools > Clip” setting the shapefile as the input layer and the overlap polygon as the overlay layer:

  1. Create a raster layer from the Clipped shapefile using the “Rasterize” tool from the toolbox setting the Output Values to “data / no-data” and Cellsize to 0.0001 for a better resolution:

  1. Right-click on the raster layer and select “Export > Save As…”. In the new window select the “Golden Software Binary Grid (.grd)” from the dropdown format menu. In the extent menu, select the “Calculate from Layer” option and then click on the “Clipped” layer:

  1. You can now import the raster from your study area into R and plot it to check whether the resolution (Cellsize) chosen is good enough:
library(raster)
LakeMacquarie <- raster("Lake_Macquarie.grd") # Import the raster file exported from QGIS
plot(LakeMacquarie) # Plot raster from the study area
If the river channels look distorted go back to step 4 and choose a lower Cellsize. Please be aware* that increasing the raster resolution too much will require higher computational costs and may cause R to crash.

If the river channels look distorted go back to step 4 and choose a lower Cellsize. Please be aware* that increasing the raster resolution too much will require higher computational costs and may cause R to crash.

Now that your acoustic detections have been filtered and that you have a raster file with good resolution from your study area (exported into your working directory), you are all set to get started with RSP.

2. Estimation of shortest in-water paths

The runRSP() function is used to recreate the shortest paths between pairs of acoustic detections. The detection data, station coordinates and the group of each fish is passed on to RSP automatically by actel through the argument input. You must also include the name of the .grid file you created above in the argument base.raster. e.g.:

library(RSP)
rsp.results <- runRSP(input = filtered_data, base.raster = "Lake_Macquarie.grd")

A transition layer object is calculated using the raster file to estimate these paths exclusively in the water. The raster is automatically imported during the analysis through the argument base.raster = "name_of_your_file.grd". Because this step can take some time depending on the size of your study area and the size of the raster cells, the transition layer will be saved for future re-analysis in case the detection data changes. When running the analysis again the following message is shown:

M: Reusing transition layer calculated on 2019-11-30 13:48:19.
   If you want to calculate a new transition layer, run rmTransition() before re-starting the analysis.

The detection ranges of each listening station are also taken into account in the runRSP(). These will be used as the location errors for the dBBMM when calculating UD areas. A ‘Range’ column can be included in the spatial.csv file for specifying the detection ranges (in meters) for each acoustic station if these are known. If the ‘Range’ column is not found, a default detection range of 500 m is automatically considered for each receiver with the warning:

Warning: Could not find a 'Range' column in the spatial data; assuming a range of 500 metres for each receiver.

Note: - The ‘Range’ column must already be present in the spatial.csv file when you run the explore() function for it to be incorporated in the analysis.

While animals move between a pair of consecutive acoustic detections there is some uncertainty regarding the trajectory taken, which increases proportionally to the time taken to from one place to another. Consecutive detections longer than 24 hours apart are thus broken by the runRSP() into separate ‘tracks’. This avoids the estimation of unrealistic behaviour when the animals do not get detected in any array for exceedingly long periods of time. Detections that occur totally isolated (e.g. more than 24-h before or after any other detection) are automatically excluded from analysis. The runRSP() will return the percentage of raw detections that can be used for refining the shortest paths when the analysis is finished:

M: Percentage of detections valid for RSP: 99.8%

Pairs of detections can occur either at the same receiver or at different receivers. For consecutive detections on different receivers, estimated positions are added according to a fixed distance argument in meters (250 m by default). On the other hand, if a fish is detected consecutively at the same station (with a time interval greater than the stipulated at the time.lapse argument), then estimated positions are added at that receiver location, over intervals of approximately [time.lapse] minutes. E.g. if a fish is detected at a station twice with a 22 minute interval, and time.lapse is set to 10, two estimated detections will be included.

While moving away from the first detection, the position errors gradually increase for each estimated position at a 5% rate of the distance. When the animal reaches half of the elapsed time/distance between the first and the second detection, the errors of estimated positions now gradually decrease as it approaches the second receiver where it got detected. This principle is used for both pairs of detections on different receivers, and for consecutive detections at the same station when the time difference is longer than time.lapse.

A: consecutive detections on the save receiver; B: consecutive detections on different receivers.

A: consecutive detections on the save receiver; B: consecutive detections on different receivers.

Please note how the distances between consecutive RSP positions vary around the distance argument (250 m by default) as they depend on the estuary shape and the shortest distances between receivers. It is also possible to observe how the errors of the estimated positions gradually increase/decrease as the animals move between receivers.

Please note how the distances between consecutive RSP positions vary around the distance argument (250 m by default) as they depend on the estuary shape and the shortest distances between receivers. It is also possible to observe how the errors of the estimated positions gradually increase/decrease as the animals move between receivers.

The dynamic Brownian Bridge Movement Model accounts for the speed at which animals move between consecutive detections to expand/contract the UD areas. Consequently, depending on your array configuration, estuary shape and species being tracked, you may find useful to adjust the distance and time.lapse arguments for recreating the most plausible movement patterns of the monitored animals.

2.1. Exploring the RSP results

Here are some examples of the runRSP() output:

  1. In the $tracks object you can find metadata, stored individually for each tracked transmitter, on the identified tracks (Track) and their corresponding number of total acoustic detections (original.n), duration in hours (Timespan), and their corresponding validity (Valid):
Track original.n First.time Last.time Timespan Valid
Track_01 3 2018-02-11 20:27:37 2018-02-11 20:29:35 0.03 hours TRUE
Track_02 2 2018-02-20 10:54:54 2018-02-20 10:56:07 0.02 hours TRUE
Track_03 103 2018-03-07 00:41:10 2018-03-07 08:20:02 7.64 hours TRUE
Track_04 22 2018-03-17 13:07:43 2018-03-17 13:36:42 0.48 hours TRUE
Track_05 1 2018-04-04 12:47:05 2018-04-04 12:47:05 0.00 hours FALSE
Track_06 2 2018-04-18 08:41:11 2018-04-18 08:48:47 0.12 hours TRUE
Track_07 3 2018-04-20 09:30:02 2018-04-20 09:33:55 0.06 hours TRUE
Track_08 7 2018-04-23 05:10:47 2018-04-23 08:43:45 3.54 hours TRUE
Track_09 22 2018-04-24 11:40:56 2018-04-26 01:00:13 37.32 hours TRUE
Track_10 5 2018-08-20 11:56:47 2018-08-20 12:06:51 0.16 hours TRUE
Track_11 2 2018-08-21 14:33:30 2018-08-21 14:42:52 0.156 hours TRUE
Track_12 2 2018-08-22 16:04:24 2018-08-22 16:05:44 0.02 hours TRUE
Track_13 1 2018-08-23 19:21:20 2018-08-23 19:21:20 0.00 hours FALSE

Only the valid tracks are used by RSP to recreate the shortest in-water paths of tracked animals. The tracking data can be retrieved from the list $detections in which data is saved individually for each trasmitter.

  1. For consecutive detections on the same receiver:
Timestamp Receiver Transmitter Error Longitude Latitude Position Track
2018-03-07 00:43:49 125449 R64K-4075 500 9.380188 56.5716 Receiver Track_3
2018-03-07 00:53:07 NA R64K-4075 512.5 9.380188 56.5716 RSP Track_3
2018-03-07 01:02:26 NA R64K-4075 512.5 9.380188 56.5716 RSP Track_3
2018-03-07 01:11:45 125449 R64K-4075 550 9.380188 56.5716 Receiver Track_3

The Position column in this dataset identifies the two consecutive acoustic detections (Receiver) from this animal. We can notice that they occurred on the same Receiver (125449): the first on 2018-03-07 00:43:49 and the second on 2018-03-07 01:11:45 (approximately 30 minutes from each other). Because this time difference is longer than the default time.lapse (10 minutes), the runRSP() estimated the intermediate positions (RSP) by repeating the receiver Longitude and Latitude and changing the Error parameter at a rate of 5% from the default distance argument (250 meters = 12.5 meters).

  1. For consecutive detections on different receivers:
Timestamp Receiver Transmitter Error Longitude Latitude Position Track
2018-04-27 05:27:10 100474 R64K-4125 500 9.921725 57.05595 Receiver Track_5
2018-04-27 05:35:17 NA R64K-4125 512.5 9.928500 57.05450 RSP Track_5
2018-04-27 05:43:24 NA R64K-4125 525 9.935500 57.05350 RSP Track_5
2018-04-27 05:51:32 NA R64K-4125 537.5 9.943500 57.05450 RSP Track_5
2018-04-27 05:59:39 NA R64K-4125 550 9.949500 57.05650 RSP Track_5
2018-04-27 06:07:47 NA R64K-4125 562.5 9.955500 57.05850 RSP Track_5
2018-04-27 06:15:54 NA R64K-4125 575 9.960500 57.06150 RSP Track_5
2018-04-27 06:24:01 NA R64K-4125 562.5 9.964500 57.06550 RSP Track_5
2018-04-27 06:32:09 NA R64K-4125 550 9.968500 57.06850 RSP Track_5
2018-04-27 06:40:16 NA R64K-4125 537.5 9.975500 57.07050 RSP Track_5
2018-04-27 06:48:24 NA R64K-4125 525 9.981500 57.07250 RSP Track_5
2018-04-27 06:56:31 NA R64K-4125 512.5 9.986500 57.07450 RSP Track_5
2018-04-27 07:04:39 107527 R64K-4125 500 9.992500 57.07650 Receiver Track_5

Here the animal was detected first at the Receiver 100474 on 2018-04-27 05:27:10, and then at the Receiver 107527 on 2018-04-27 07:04:39. The runRSP() now calculated the shortest in-water path between receivers, and we can see how the Error of added locations increased up to half-way, (575 meters on 2018-04-27 06:15:54), and then decreased back to 500 as the track approached the second receiver.

2.2. Visualizing RSP outputs

We can use plotDist() to compare the total distances travelled by each animal calculated using only the receiver locations and or also including the RSP estimations:

The plotDetec() shows the total number of receiver and estimated positions for each tracked animal:

You can also plot the tracks from a particular animal using plotRSP():

plotRSP(input = rsp.results, tag = "R64K-4075", display = "Receiver", type = "lines") 
plotRSP(input = rsp.results, tag = "R64K-4075", display = "RSP", type = "lines") 
plotRSP(input = rsp.results, tag = "R64K-4138", display = "Receiver", type = "lines") 
plotRSP(input = rsp.results, tag = "R64K-4138", display = "RSP", type = "lines")

You can also set display = "Both" to plot both track options on a single plot.

3. Calculating utilization distribution areas and space-use overlaps

After estimating the in-water shortest paths, we can now use the output from runRSP() to calculate UD areas with the dynBBMM() function. Here you will need to know the UTM zone of your study site and specify it using the argument UTM.zone. By default, the analysis will run for all transmitters detected, but you can determine also which transmitters you would like to include using tags. As mentioned before, UD areas can be calculated with either of the following temporal resolutions:

3.1. Total dynamic Brownian Bridge Movement Model (group dBBMM)

This option calculates a series of dBBMM for each animal track from all the groups monitored. The breaks argument defines for which contours the areas of use should be calculated, which by default are the 50% and 95% (i.e. breaks = c(.5, .95)).

Track quality checks are performed to ensure that only good tracks which allow the dBBMM to converge are included in the analysis. This is an example of the returned messages from dynBBMM():

dbbmm.results <- dynBBMM(input = rsp.results, UTM.zone = 56, breaks = c(0.5, 0.95))

M: Preparing data to apply dBBMM.
M: No specific transmitters selected. All the data will be used for analysis.
Warning: 7 track(s) in group R64K-4075 have less than eight detections and will not be used.
Warning: 1 track(s) in group R64K-4075 are shorter than 30 minutes and will not be used.
Warning: 2 individual detections were removed in group R64K-4125 due to simultaneous detections at two receivers.
Warning: 1 track(s) in group R64K-4125 have less than eight detections and will not be used.
Warning: 1 track(s) in group R64K-4128 have less than eight detections and will not be used.
Warning: 2 track(s) in group R64K-4128 are shorter than 30 minutes and will not be used.
Warning: 6 track(s) in group R64K-4138 have less than eight detections and will not be used.
M: In total, 93 detections were excluded as they failed the track quality checks.

After calculating UDs, the land areas are excluded so that the final results represent only in-water areas of use. The overall overlap between each group monitored is also calculated.

M: Subtracting land areas from output.
M: Calculating overlaps between groups.
M: Storing final results.

The results of the dBBMM are saved in the $track.areas object, as a list of data frames for each group analysed:

Track Start Stop Area.5 Area.95 Time.lapse.min
R64K-4125_Track_2 2018-04-21 13:13:24 2018-04-23 09:09:34 457 4021 2636.16667
R64K-4125_Track_3 2018-04-25 11:44:05 2018-04-28 14:10:14 2016 28434 4466.15000
R64K-4125_Track_4 2018-08-23 12:16:57 2018-08-24 11:10:14 2617 4693 1373.28333
R64K-4125_Track_5 2018-08-25 13:37:35 2018-08-25 15:08:31 53 196 90.93333
R64K-4125_Track_6 2018-08-27 08:44:04 2018-08-27 15:38:32 50 175 414.46667
R64K-4125_Track_8 2018-08-30 17:26:41 2018-08-31 15:43:46 80 418 1337.08333

Each Track is named after the transmitter and the corresponding track name and both Start and Stop timestamps are stored. The areas of use (by default Area.5 = 50% and Area.95 = 95%) are saved in squared meters, together with the respective elapsed times in minutes (Time.lapse.min). Note that in this example for the animal R64K-4125 the tracks 1 and 7 failed the quality checks and thus not included in the analysis.

You can use plotContours() to visualize any of the dBBMM calculated by specifying the group and track you want to plot:

plotContours(input = dbbmm.results, group = "R64K-4138", track = "R64K-4138_Track_10")
plotContours(input = dbbmm.results, group = "R64K-4125", track = "R64K-4125_Track_4")
The levels argument can be used to specify which contours to plot. By default, the 25%, 50%, 75%, 95% and 99% contours are returned.

The levels argument can be used to specify which contours to plot. By default, the 25%, 50%, 75%, 95% and 99% contours are returned.

3.2. Total space-use overlap

Now that we calculated the areas of space-use within our study area for each group monitored, we can investigate the ammount of overall overlap between them stored in the $overlap.areas object:

dbbmm.results$overlap.areas$`0.5`$absolute
R64K-4075 R64K-4125 R64K-4128 R64K-4138
R64K-4075 NA 1716 1792 3606
R64K-4125 1716 NA 1769 3551
R64K-4128 1792 1769 NA 2635
R64K-4138 3606 3551 2635 NA 
dbbmm.results$overlap.areas$`0.5`$percentage
R64K-4075 R64K-4125 R64K-4128 R64K-4138
R64K-4075 NA 0.3622546 0.6775047 0.6049321
R64K-4125 0.3622546 NA 0.6688091 0.7496306
R64K-4128 0.6775047 0.6688091 NA 0.9962193
R64K-4138 0.6049321 0.7496306 0.9962193 NA 
dbbmm.results$overlap.areas$`0.95`$absolute
R64K-4075 R64K-4125 R64K-4128 R64K-4138
R64K-4075 NA 12851 4893 17480
R64K-4125 12851 NA 5115 29210
R64K-4128 4893 5115 NA 6839
R64K-4138 17480 29210 6839 NA 
dbbmm.results$overlap.areas$`0.95`$percentage
R64K-4075 R64K-4125 R64K-4128 R64K-4138
R64K-4075 NA 0.7247349 0.7133693 0.9857884
R64K-4125 0.7247349 NA 0.7457355 0.9064672
R64K-4128 0.7133693 0.7457355 NA 0.9970841
R64K-4138 0.9857884 0.9064672 0.9970841 NA

Please note the overlaps are calculated for the contours defined by breaks in dynBBMM(), and returned both in absolute values (squared meters) and percentage matrices. For example, we can see in the last table that R64K-4128 and R64K-4138 were the groups with higher overall overlap of 99.71% at the 95% level, whereas R64K-4075 and R64K-4128 had a smaller overlap of 71.34% at the 95% contour. To see exactly where space use overlaps occurred you can use plotOverlap():

plotOverlap(input = dbbmm.results, stations = FALSE, level = .95, store = TRUE)

3.3. Fine-scale dynamic Brownian Bridge Movement Model (timeslot dBBMM)

dBBMMs can also be calculated according to a moving temporal window. This allows investigating how the space-use overlap between the different groups varied during the study period. It is useful for assessing the influence of environmental parameters upon space-use of different groups tracked within the study area. The same dynBBMM() function is used, but here the argument timeframe has to be defined in hours as the temporal window. The total tracking period will be divided into timeslots, and dBBMMs calculated for each group monitored (for each timeslot). Overlapping areas are now calculated for each timeslot and the corresponding metadata stored in the $timeslots object:

time.dbbmm.results <- dynBBMM(input = rsp.results, UTM.zone = 56, breaks = c(0.5, 0.95), timeframe = 24) # 24-h timeslots
time.dbbmm.results$timeslots[400:410, ]
slot start stop Bream Luderick Tarwhine
400 2014-10-06 00:00:00 2014-10-07 00:00:00 FALSE FALSE FALSE
401 2014-10-07 00:00:00 2014-10-08 00:00:00 FALSE FALSE FALSE
402 2014-10-08 00:00:00 2014-10-09 00:00:00 FALSE FALSE FALSE
403 2014-10-09 00:00:00 2014-10-10 00:00:00 FALSE FALSE FALSE
404 2014-10-10 00:00:00 2014-10-11 00:00:00 TRUE FALSE TRUE
405 2014-10-11 00:00:00 2014-10-12 00:00:00 FALSE FALSE TRUE
406 2014-10-12 00:00:00 2014-10-13 00:00:00 FALSE FALSE TRUE
407 2014-10-13 00:00:00 2014-10-14 00:00:00 FALSE FALSE TRUE
408 2014-10-14 00:00:00 2014-10-15 00:00:00 FALSE FALSE TRUE
409 2014-10-15 00:00:00 2014-10-16 00:00:00 FALSE FALSE TRUE
410 2014-10-16 00:00:00 2014-10-17 00:00:00 FALSE FALSE TRUE

In the example above (daily timeslots set using timeframe = 24) we can notice that in the four first timeslot none of the three groups monitored were detected, whereas both Bream and Tarwhine groups got detected in slot 404 and from slot 405 to slot 410 only Tarwhine were detected.

The $track.areas object for each tracked group will now have a first column named Slot, which now identifies the timeslot for each of the dBBMM calculated:

time.dbbmm.results$track.areas$Bream[24:30, ]
Slot Track Start Stop Area.5 Area.95 Time.lapse.min
404 A69-9004-485_Track_01 2014-10-10 10:08:08 2014-10-10 10:38:15 289 1245 30.11667
411 A69-9004-485_Track_02 2014-10-17 12:41:35 2014-10-17 23:59:03 635 3489 677.46667
412 A69-9004-485_Track_02 2014-10-18 00:09:01 2014-10-18 23:26:49 670 4518 1397.80000
413 A69-9004-485_Track_02 2014-10-19 00:18:55 2014-10-19 06:45:37 231 1062 386.70000
425 A69-9004-485_Track_04 2014-10-31 10:24:47 2014-10-31 23:51:01 685 2731 806.23333
426 A69-9004-485_Track_04 2014-11-01 00:00:59 2014-11-01 23:58:59 1158 5344 1438.00000
427 A69-9004-485_Track_04 2014-11-02 00:08:18 2014-11-02 23:53:08 624 2515 1424.83333

Here we can see that the Bream A69-9004-485 was detected consecutivelly between 2014-10-17 12:41:35 and 2014-10-19 06:45:37 (Track_02) and again between 2014-10-31 10:24:47 and 2014-11-02 23:53:08 (Track_04).

The following command line can help you assess if any other group got detected during a particular timeslot:

> time.dbbmm.results$timeslots[404, ]
    slot               start                stop Bream Luderick Tarwhine
404  404 2014-10-10 00:00:00 2014-10-11 00:00:00  TRUE    FALSE     TRUE

Yes, both Bream and Tarwhine were detected between 2014-10-10 00:00:00 and 2014-10-11 00:00:00 (timeslot 404). We can now inspect whether the two groups overlapped or not:

> time.dbbmm.results$overlap.areas$`0.95`$percentage$`404`
             Bream Luderick  Tarwhine
Bream           NA       NA 0.9799197
Luderick        NA       NA        NA
Tarwhine 0.9799197       NA        NA

This shows that the two groups had an overlap of 97.99% at the 95% dBBMM contour during this particular timeslot (or in this case, day). We can now see exactly where the overlap occurred by plotting the space use models and the overlap contours using:

plotContours(input = time.dbbmm.results, group = "Bream", track = "A69-9004-485_Track_01", main = "A69-9004-485 (Bream)", stations = TRUE, timeslot = 404)
plotContours(input = time.dbbmm.results, group = "Tarwhine", track = "A69-9004-489_Track_1", main = "A69-9004-489 (Tarwhine)", stations = TRUE, timeslot = 404)
Setting the argument stations = TRUE will include the locations of acoustic receivers in the plot.

Setting the argument stations = TRUE will include the locations of acoustic receivers in the plot.

Here we can see how these two transmitters really occurred in a similar region during timeslot = 404. To inspect for the total space use areas at group level, and the exact overlap in space and time we can use plotOverlap():

> plotOverlap(input = dbbmm_time_24, level = .95, store = TRUE, stations = FALSE, timeslot = 404, main = "2014-10-10")
M: No overlap found between 'Bream' and 'Luderick'.
M: No overlap found between 'Luderick' and 'Tarwhine'.